System and method for hypertension monitoring

Abstract

A wearable device can be used for hypertension monitoring. The wearable device can include motion sensors and optical sensors. Data from these sensors can be processed in the wearable device and / or by another device in communication with the wearable device to provide early screening for undiagnosed hypertension. If the screening estimates that the user has undiagnosed hypertension, the user can be notified to seek proper hypertension diagnosis. The hypertension monitoring can include a first phase to estimate one or more short-term hypertension scores or parameters. The hypertension monitoring can also include a second phase to estimate a long-term hypertension score using the accumulated short-term scores / parameters to estimate hypertension.

Description

[0001] Cross-references to related applications

[0002] This application claims the benefits of U.S. Provisional Application No. 63 / 033,802, filed June 2, 2020, and U.S. Provisional Application No. 63 / 146,536, filed February 5, 2021, the contents of which are incorporated herein by reference in their entirety for all purposes. Technical Field

[0003] This disclosure relates in general to systems and methods for monitoring hypertension, and more specifically, to hypertension monitoring using wearable devices. Background Technology

[0004] Without proper diagnosis and treatment, high blood pressure (high blood pressure) can increase the risk of health problems such as stroke and heart attack. High blood pressure often goes undetected because it may not cause symptoms for months or even years. However, even without symptoms, high blood pressure can damage the heart and blood vessels. Therefore, providing users with information about high blood pressure can help improve their health. Summary of the Invention

[0005] This disclosure relates to systems and methods for monitoring hypertension using wearable devices. Wearable devices may include motion and / or orientation sensors (e.g., accelerometers, gyroscopes, inertial measurement units (IMUs), etc.) and optical sensors. Data from the sensors may be processed within the wearable device and / or by another device communicating with the wearable device to provide early screening for undiagnosed hypertension. If the screening estimates undiagnosed hypertension in the user, the user may be notified to seek a proper hypertension diagnosis. Hypertension monitoring may include a first phase to estimate one or more short-term hypertension scores or parameters. Short-term hypertension scores / parameters may be correlated with blood pressure. In some examples, short-term hypertension scores / parameters may include a systolic blood pressure score (or parameter) and a diastolic blood pressure score (or parameter). Hypertension monitoring may also include a second phase to estimate a long-term hypertension score using the accumulated short-term scores / parameters (e.g., short-term hypertension scores / parameters for a threshold time period or a threshold number of short-term hypertension scores / parameters) to estimate hypertension. Attached Figure Description

[0006] Figures 1A to 1B An exemplary system for monitoring hypertension, according to an example of this disclosure, is shown.

[0007] Figure 2 An exemplary block diagram of hypertension monitoring according to an example of this disclosure is shown.

[0008] Figure 3An exemplary process for hypertension monitoring is shown according to an example of this disclosure.

[0009] Figures 4A to 4B An exemplary block diagram of a short-term hypertension score generator according to an example of this disclosure is shown.

[0010] Figures 5A to 5C An exemplary block diagram of a long-term hypertension score generator according to an example of this disclosure is shown. Detailed Implementation

[0011] In the following description of the examples, reference will be made to the accompanying drawings, which form part of the following description, and specific examples that can be implemented are shown by way of example in the drawings. It should be understood that other examples and structural changes may be used without departing from the scope of the disclosed examples.

[0012] This disclosure relates to systems and methods for monitoring hypertension using wearable devices. Wearable devices may include motion and / or orientation sensors (e.g., accelerometers, gyroscopes, inertial measurement units (IMUs), etc.) and optical sensors. Data from the sensors may be processed within the wearable device and / or by another device communicating with the wearable device to provide early screening for undiagnosed hypertension. If the screening estimates undiagnosed hypertension in the user, the user may be notified to seek a proper hypertension diagnosis.

[0013] Hypertension monitoring may include a first phase to estimate one or more short-term hypertension scores or parameters. Short-term hypertension scores / parameters may be correlated with blood pressure. In some examples, short-term hypertension scores / parameters may include systolic blood pressure scores (or parameters) and diastolic blood pressure scores (or parameters). Hypertension monitoring may also include a second phase to estimate long-term hypertension scores using the accumulated short-term scores / parameters (e.g., short-term hypertension scores / parameters for a threshold time period or a threshold number) to estimate hypertension.

[0014] As used herein, a "short-term" hypertension score / parameter can represent a hypertension score / parameter calculated based on segments of input data from one or more sensors, each segment corresponding to a first time period (e.g., 30 seconds, 1 minute, 2 minutes, 5 minutes, etc.). The short-term hypertension score / parameter can be correlated with blood pressure in segments (e.g., including data acquired within the first time period). As used herein, a "long-term" hypertension score can represent a hypertension score calculated based on input data acquired within a second time period (e.g., several days, a week, several weeks, a month, etc.), which can be correlated with blood pressure in the second time period. Therefore, "short-term" and "long-term" reflect the relative difference between the first and second time periods. The second time period used for the "long-term" hypertension score can be several orders of magnitude longer than the first time period used for the "short-term" hypertension score / parameter.

[0015] Figures 1A to 1B An exemplary system for monitoring hypertension, according to an example of this disclosure, is shown. The system may include one or more sensors and processing circuitry to estimate hypertension over a period of time using data from the one or more sensors. In some examples, the system may be implemented in a wearable device (e.g., wearable device 100). In some examples, the system may be implemented in more than one device (e.g., wearable device 100 and a second device communicating with wearable device 100).

[0016] Figure 1A An exemplary wearable device 100 is shown that can be attached to a user using a strip 146 or other fasteners. The wearable device 100 may include one or more sensors for estimating hypertension over a period of time using data from one or more sensors, and may optionally include a touchscreen 128 to display the results of hypertension monitoring as described herein.

[0017] Figure 1B An exemplary block diagram of the architecture of a wearable device 100 for monitoring high blood pressure, according to an example of this disclosure, is shown. Figure 1BAs shown, wearable device 100 may include one or more sensors. For example, wearable device 100 may optionally include an optical sensor comprising one or more light emitters 102 (e.g., one or more light-emitting diodes (LEDs)) and one or more optical sensors 104 (e.g., one or more photodetectors / photodiodes). The one or more light emitters may generate light in the range corresponding to infrared (IR), green, amber, blue, and / or red light, among others. The optical sensors may be used to emit light into the user's skin 114 and detect the reflection of light back from the skin. The optical sensor measurements may represent a time-domain photoplethysmography (PPG) signal. The optical sensor measurements may be converted into a digital signal via an analog-to-digital converter (ADC) 105b for processing. In some examples, the optical sensors and the processing of the light signal by one or more processors 108 may be used for various functions, including estimating physiological characteristics (e.g., heart rate, arterial oxygen saturation, etc.), monitoring physiological conditions (e.g., hypertension), and / or detecting contact with the user (e.g., on / off wrist detection).

[0018] In some examples, processing of the optical signal by one or more processors 108 may include identifying cardiac cycles (pulses) in the optical signal from an optical sensor. For example, processing of the optical signal may include identifying one or more features (e.g., systolic peak, diastolic notch, diastolic peak, etc.) of the cardiac cycle in the PPG signal. One or more features may be used to identify each cardiac cycle (e.g., those not disrupted by motion artifacts) and the temporal location of the cardiac cycle (e.g., timing corresponding to one of the features). Additionally, processing of the optical signal by one or more processors 108 may include (e.g., based on the morphology of the PPG signal) calculating a confidence parameter associated with each cardiac cycle. In some examples, processing of the optical signal by one or more processors 108 may include using one or more features of the cardiac cycle to identify qualified cardiac cycles (qualified pulses) in the optical signal, wherein the confidence parameter meets one or more qualification criteria. In some examples, the cardiac cycle may be qualified when the confidence parameter is above a threshold and unqualified when the confidence parameter is below a threshold.

[0019] One or more sensors may include motion and / or orientation sensors, such as accelerometers, gyroscopes, inertial measurement units (IMUs), etc. For example, wearable device 100 may include accelerometer 106, which may be a multi-channel accelerometer (e.g., a 3-axis accelerometer). As described in more detail herein, motion and / or orientation sensors may be used for hypertension monitoring. In some examples, motion and / or orientation information may be used to provide an indication of motion artifacts and / or user posture that may affect (e.g., disrupt) some samples of the PPG signal. Additionally or alternatively, motion and / or orientation data may also carry information about the heartbeat, and this information (and its timing relative to the heartbeat in the PPG signal) may be used to estimate a hypertension score / parameter as described herein. Measurements from accelerometer 106 may be converted into digital signals for processing via ADC 105a.

[0020] Wearable device 100 may also optionally include other sensors, including but not limited to photothermal sensors, magnetometers, barometers, compasses, proximity sensors, cameras, ambient light sensors, thermometers, GPS sensors, and various system sensors capable of sensing remaining battery life, power consumption, processor speed, CPU load, etc. While a variety of sensors are described, it should be understood that fewer, more, or different sensors may be used.

[0021] Data acquired from one or more sensors (e.g., motion data, optical data, etc.) may be stored in the memory of the wearable device 100. For example, the wearable device 100 may include a data buffer (or other volatile or non-volatile memory or storage device) to temporarily (or permanently) store data from the sensors for processing by processing circuitry. In some examples, the volatile or non-volatile memory or storage device may be used to store processed data (e.g., filtered data, short-term hypertension scores or parameters, long-term hypertension scores, etc.) for further processing or for storing and / or displaying hypertension monitoring results. In some examples, the volatile or non-volatile memory or storage device may be used to store processed data, referred herein as pulse data indicating the location of a qualified pulse or indicating the location of a pulse, and a confidence parameter associated with the location of the pulse. Additionally or alternatively, volatile or non-volatile memory or storage device may be used to store processed data referred to herein as extracted feature data, which includes features extracted from optical data on a per-pulse basis (optionally, the features extracted per pulse across some or all pulses in the input segment, or otherwise aggregated) in the input segment.

[0022] Wearable device 100 may also include processing circuitry to perform the various processes described herein, including generating hypertension scores / parameters and estimating hypertension. The processing circuitry may include one or more processors 108. These processors may include a digital signal processor (DSP) 109, a microprocessor, a central processing unit (CPU), a programmable logic device (PLD), a field-programmable logic array (FPGA), etc.

[0023] In some examples, some processing may be performed by a peripheral device 118 that communicates with the wearable device. The peripheral device 118 may be a smartphone, media player, tablet computer, desktop computer, laptop computer, data server, cloud storage service, or any other portable or non-portable electronic computing device (including the second wearable device). The wearable device 100 may also include communication circuitry 110 to be communicatively coupled to the peripheral device 118 via a wired or wireless communication link 124. For example, the communication circuitry 110 may include circuitry for one or more wireless communication protocols, including cellular, Bluetooth, Wi-Fi, etc.

[0024] In some examples, wearable device 100 may include a touchscreen 128 to display hypertension monitoring results (e.g., displaying a notification seeking a medical diagnosis) and / or receive input from the user. In some examples, touchscreen 128 may be replaced by a non-touch-sensitive display, or the touch and / or display functionality may be implemented in another device. In some examples, wearable device 100 may include a microphone / speaker 122 for audio input / output functionality, haptic circuitry to provide haptic feedback to the user, and / or other sensors and input / output devices. Wearable device 100 may also include an energy storage device (e.g., a battery) to power the components of wearable device 100.

[0025] One or more processors 108 (also referred to herein as processing circuitry) may be connected to program storage device 111 and may be configured (programmed) to execute instructions stored in program storage device 111 (e.g., a non-transitory computer-readable storage medium). For example, the processing circuitry may provide control and data signals to generate display images on touchscreen 128, such as display images of a user interface (UI), optionally including results of hypertension monitoring. The processing circuitry may also receive touch input from touchscreen 128. Touch input can be used by a computer program stored in program storage device 111 to perform actions, including but not limited to: moving objects such as cursors or pointers, scrolling or panning, adjusting control settings, opening files or documents, viewing menus, making selections, executing instructions, operating peripherals connected to the host device, answering telephone calls, making telephone calls, terminating telephone calls, changing volume or audio settings, storing information related to telephone communication (such as addresses, frequently dialed numbers, received calls, missed calls), logging onto a computer or computer network, allowing authorized individuals access to restricted areas of a computer or computer network, loading user profiles associated with the user's preferred computer desktop layout, allowing access to web page content, launching specific programs, encrypting or decrypting messages, etc. The processing circuitry can also perform additional functions that may not be related to touch processing and display. In some examples, the processing circuitry can perform some of the signal processing functions described herein (e.g., hypertension monitoring / scoring).

[0026] It should be noted that one or more functions described herein, including hypertension monitoring, can be executed by firmware stored in memory or by instructions stored in program storage device 111 and executed by processing circuitry. The firmware may also be stored and / or delivered to any non-transitory computer-readable storage medium for use or in conjunction with an instruction execution system, apparatus, or device, such as a computer-based system, a processor-based system, or other system that can retrieve and execute instructions from and from the instruction execution system, apparatus, or device. In the context of this document, "non-transitory computer-readable storage medium" can be any medium (excluding signals) that can contain or store programs for use or in conjunction with an instruction execution system, apparatus, or device. Computer-readable storage media may include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices; portable computer disks (magnetic); random access memory (RAM) (magnetic); read-only memory (ROM) (magnetic); erasable programmable read-only memory (EPROM) (magnetic); or flash memory such as compact flash cards, secure digital cards, USB memory devices, memory sticks, etc.

[0027] This firmware can also be propagated in any transmission medium for use or in conjunction with an instruction execution system, apparatus, or device, such as a computer-based system, a processor-based system, or other system capable of retrieving and executing instructions from and with an instruction execution system, apparatus, or device. In the context of this document, "transmission medium" can be any medium through which a program can be transmitted, propagated, or transferred for use or in conjunction with an instruction execution system, apparatus, or device. Transmission media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, or infrared wired or wireless transmission media.

[0028] It is obvious that Figure 1B The architecture shown is merely an example, and wearable devices may have more or fewer components than those shown, or different component configurations. Figure 1B The various components shown can be implemented in hardware, software, firmware, or any combination thereof (including one or more signal processing and / or application-specific integrated circuits). Additionally, Figure 1B The components shown may be included in a single device or distributed among multiple devices.

[0029] Figure 2 An exemplary block diagram of hypertension monitoring according to an example of this disclosure is shown. Block diagram 200 may include processing circuitry (e.g., corresponding to...). Figure 1B One or more processors 108 and / or DSP 109 are used to calculate hypertension scores and / or parameters. In some examples, the processing circuitry may include a short-term hypertension score generator 205 (first stage) and a long-term hypertension score generator 215 (second stage). The block diagram may also include a memory 210 that can store short-term hypertension scores / parameters generated by the short-term hypertension score generator 205 and can be accessed by the long-term hypertension score generator 215.

[0030] The short-term hypertension score generator 205 can receive data from one or more sensors as input. The data may include optical data from an optical sensor (e.g., a PPG signal) and motion data from a motion sensor (e.g., a triaxial accelerometer). Both optical and motion data may be captured in parallel within segments of a first time period (e.g., 30 seconds, 1 minute, 2 minutes, 5 minutes, etc.). In some examples, the input to the short-term hypertension score generator 205 may also include pulse data indicating the location of a qualified pulse (or the location of the pulse and a confidence parameter associated with the location of the qualified pulse). The acquisition of optical and motion data (and / or the processing of optical and / or motion data for generating pulse data) may be part of background processing performed without a prior user request for data acquisition. Additionally or alternatively, in some examples, the acquisition of optical and motion data (and / or the processing of optical and / or motion data for generating pulse data) may be in response to a user request (e.g., a user request to measure heart rate using optical and motion sensors). In some examples, background processing for acquiring optical and motion data (and / or processing optical and / or motion data to generate pulse data) can be performed continuously, periodically (e.g., integer times per hour or per day), within a threshold time period after the last measurement of the optical / motion data, or in response to various triggers. In some examples, the frequency of background processing for acquiring optical and motion data (and / or processing optical and / or motion data to generate pulse data) can be limited as a function of the total power allocated to the background processing.

[0031] Short-term hypertension score generator 205 can process segments of optical and accelerometer data (and optionally, pulse data) to generate short-term hypertension scores / parameters that can be correlated with blood pressure. In some examples, the short-term hypertension score / parameter may include estimated systolic and diastolic blood pressure (or systolic and diastolic scores correlated with systolic and diastolic blood pressure) for each segment. In some examples, the short-term hypertension score / parameter may include multiple parameters (e.g., corresponding to features) extracted from the input data, rather than a single score for each segment. In some examples, the short-term hypertension score / parameter may include an estimated overall hypertension score for each segment (without decomposing the systolic and diastolic scores). In some examples, the short-term hypertension score / parameter may be estimated for sub-segments within a segment (e.g., on a per-pulse basis). In some examples, the short-term hypertension score / parameter for each segment may be stored in memory 210.

[0032] In some examples, once a segment (e.g., the first time segment) of data is acquired (e.g., in response to acquiring sufficient optical and motion data in a first time period to perform a short-term hypertension score), the optical and motion data (and optionally, pulse data) can be processed by the short-term hypertension score generator 205. In some examples, segments of optical and motion data (and optionally, pulse data) can be stored (e.g., stored in memory 210 or in a data buffer (not shown)) and can be processed later.

[0033] The long-term hypertension score generator 215 can process short-term hypertension scores / parameters (e.g., from memory 210) and can generate a long-term hypertension score that can be correlated with blood pressure and / or hypertension estimates. The long-term hypertension score can be estimated using the aggregation of short-term hypertension scores / parameters during a second time period.

[0034] Figure 3 An exemplary process for hypertension monitoring according to an example of this disclosure is shown. Process 300 may be performed by processing circuitry including processor 108 and / or DSP 109. At 305, optical and motion data may be acquired over a duration of a first time period (e.g., 30 seconds, 1 minute, 2 minutes, 5 minutes, etc.). In some examples, in addition to acquiring optical and motion data, at 308, pulse data indicating a qualified pulse for a segment may be acquired (or generated using optical and / or motion data). The acquisition of optical data segments and motion data (and optionally, pulse data) may be part of background processing. At 310, a short-term hypertension score / parameter may be generated by a short-term hypertension score generator 205 using segments of optical and motion data (and optionally, pulse data). If it is determined that the acquired short-term hypertension score / parameter is insufficient (315), the acquisition of optical and motion data at 305 (and optionally, the acquisition of pulse data at 308) and the short-term hypertension score at 310 may be repeated. Based on the determination that the acquired short-term hypertension score / parameters are sufficient (315), at 320, a long-term hypertension score can be generated by the long-term hypertension score generator 215.

[0035] In some examples, the adequacy or inadequacy of short-term hypertension scores / parameters can be determined based on a threshold number (e.g., 50, 100, 120, 250, etc.) of short-term hypertension scores / parameters corresponding to a threshold number of motion and optical data segments. In some examples, the adequacy or inadequacy of short-term hypertension scores / parameters can be based on a time period. For example, adequacy of short-term hypertension scores / parameters can be determined after a threshold time period, such as a second time period (e.g., several days, a week, several weeks, a month, etc.). In some examples, adequacy of short-term hypertension scores / parameters can be determined by having a threshold number of short-term hypertension scores / parameters for each sub-time period of the time period (e.g., for at least one segment used to generate short-term hypertension scores / parameters for each day in the second time period).

[0036] At point 325, a threshold can be applied to the long-term hypertension score / parameter. If the long-term hypertension score / parameter exceeds the threshold, the hypertension estimate can be reported to the user at point 330. For example, a notification can be displayed to indicate the possibility of undiagnosed hypertension and / or to advise the user to seek medical care for a diagnosis of hypertension. Additionally or alternatively, in some examples, the user can receive feedback including audio and / or tactile feedback regarding hypertension monitoring. In some examples, the results can be reported to a health application. In some examples, a notification can be provided to the user's doctor / medical team if authorized by the user. If the long-term hypertension score / parameter does not exceed the threshold, the results may not be notified to the user. In some examples, the hypertension monitoring process according to process 300 can be repeated (e.g., in a second time period) to continue monitoring hypertension.

[0037] The threshold used at 325 can be adjusted based on empirical data to reduce the number of false positive results (e.g., incorrect indications of hypertension) and increase the number of true positive results (e.g., true indications of hypertension). In some examples, the threshold can be adjusted to maximize the number of true positive notifications for stage II hypertension and minimize the number of notifications for non-hypertensive individuals. Although process 300 is described as generating a long-term hypertension score and using a single threshold to distinguish between hypertension estimates (and reported) and non-hypertensive (and unreported) estimates, it should be understood that in some examples, multiple thresholds can be used to distinguish between multiple levels of blood pressure. For example, the threshold can distinguish between non-hypertensive, elevated blood pressure, stage I hypertension, or stage II hypertension. In some examples, results may be reported to the user for some levels (e.g., stage I and stage II hypertension) while not reported for others (e.g., elevated blood pressure without hypertension). In some examples, a specific level may be reported as part of the hypertension estimate reported to the user and / or may be reported to the user's physician (reporting or not reporting the specific level to the user).

[0038] Figures 4A to 4B An exemplary block diagram of a short-term hypertension score generator according to an example of this disclosure is shown. Figure 4A An exemplary block diagram of a short-term hypertension score generator according to an example of this disclosure is shown. The short-term hypertension score generator 400 may correspond to... Figure 2 The short-term hypertension score generator 405 is described in some examples. In some examples, the short-term hypertension score generator 400 may include an optical data filter 405, a motion data filter 410, and machine learning processing circuitry 415. The optical data filter 405 may bandpass filter (e.g., allowing frequencies in the range of 0.1 Hz to 8 Hz or 0.5 Hz to 20 Hz) on optical data from optical sensors (e.g., PPG signals from optical sensors including one or more light emitters 102 and one or more optical sensors 104). The motion data filter 410 may bandpass filter motion data from motion sensors (e.g., a multi-axis accelerometer 106).

[0039] The machine learning processing circuit 415 can be a dual-head convolutional neural network (CNN) with self-attention. In some examples, the backbone of the CNN may include multiple convolutional layers organized into residual blocks. The CNN can transform the input temporal tensor of filtered optical and motion data (e.g., fragments of optical / motion data) to extract a set of features (short-term hypertension parameters) for the prediction head. The prediction head can then compute these features to generate systolic and diastolic hypertension scores. For example, a first prediction head output in the prediction head output could be a diastolic hypertension score 420, and a second prediction head output could be a systolic hypertension score 425. These systolic and diastolic hypertension scores can be correlated with the blood pressure of the sample. In some examples, each prediction head can be embedded with a self-attention mechanism that enables the corresponding prediction head to pay attention to the part of the feature space that is most prominent for its target (e.g., systolic or diastolic blood pressure). Both the feature representation (e.g., feature set) and the self-attention can be learned automatically from the labeled training data in an end-to-end manner. For example, training data can be acquired by measuring blood pressure with a device (e.g., a blood pressure cuff) to provide labeled systolic and diastolic blood pressure simultaneously with (or close to) measuring PPG and accelerometer data via a wearable device. In some examples, the coefficients of the CNN can be tuned to minimize the mean absolute error (MAE) between the output short-term hypertension score and the short-term systolic and diastolic blood pressure labels from the training data. In some examples, features can be stored as short-term hypertension parameters instead of calculating a short-term hypertension score. The short-term hypertension score and / or parameters can be stored in memory.

[0040] Figure 4B Another exemplary block diagram of a short-term hypertension score generator according to an example of this disclosure is shown. The short-term hypertension score generator 450 may correspond to... Figure 2 The short-term hypertension score generator 450 is described in several examples. In some examples, the short-term hypertension score generator 450 may receive filtered optical data, filtered motion data, and pulse data. In some examples, the pulse data may include information about the relative temporal position of one or more pulses and / or information about the quality of the optical data corresponding to the pulses. In some examples, the pulse data may include the position of qualified pulses (e.g., pulses with confidence parameters that meet one or more defined criteria). In some examples, the relative temporal position of the pulses may be defined by the temporal position of specific features in the morphology of the optical signal (such as features of the cardiac cycle represented in the optical signal). In some examples, the features may be a systolic peak, a diastolic notch, or a diastolic peak. In some examples, the short-term hypertension score generator 450 may include an optical data filter (e.g., similar to a reference filter). Figure 4A The optical data filter 405 is described, but... Figure 4B (not shown in the image) and / or motion data filters (e.g., similar to reference filters). Figure 4A The motion data filter 410 is described, but... Figure 4B (Not shown in the image). In some examples, the short-term hypertension score generator 450 may also receive extracted feature data (e.g., the frequency, amplitude, phase, and / or other timing characteristics of the PPG signal) extracted from optical data. In some examples, the extracted feature data may be extracted from optical data of one or more pulses that meet the same or similar eligibility criteria (e.g., pulses that meet or exceed a threshold confidence level).

[0041] The short-term hypertension score generator 450 may include preprocessing circuitry 455, machine learning processing circuitry 460, and transformation circuitry 465. Preprocessing circuitry 455 (also referred to herein as a preprocessor) may receive an input time-series tensor of filtered optical and motion data (e.g., segments of optical / motion data) and pulse data to divide the input time-series tensor of filtered optical and motion data into discrete sub-segments (also referred to herein as pulse windows). In some examples, each sub-segment / pulse window may have the same duration (e.g., 0.5 seconds, 0.75 seconds, 1 second, etc.) and may be defined relative to the temporal position of a qualifying pulse indicated by the pulse data. In some examples, the pulse window may be centered on the temporal position of the pulse indicated by the pulse data. In some examples, the pulse window may begin at the temporal position of the pulse indicated by the pulse data.

[0042] In some examples, the number of pulse windows for the input time series tensor can be limited (e.g., at 50 pulse windows, 60 pulse windows, 70 pulse windows, etc.). In some examples, the number of pulse windows can be limited such that the number of valid pulses multiplied by the duration of the pulse window is less than or equal to the duration of the input time series tensor. In some examples, pulse windows corresponding to the pulses with the highest confidence are used, and a maximum number of pulse windows higher than the one with the lowest confidence can be discarded (not for short-term hypertension scoring). In some examples, the input time series tensor is sequentially divided into pulse windows, and the partitioning of the input time series tensor can be terminated once the maximum number of pulse windows has been achieved.

[0043] In some examples, a minimum number of pulse windows may be required for short-term hypertension scoring. When the number of pulse windows is less than the minimum number (e.g., determined as part of preprocessing), the short-term hypertension score generator can bypass the short-term hypertension scoring of the input temporal tensor with less than the minimum number of pulse windows. When the number of pulse windows is equal to or greater than the minimum number (e.g., determined as part of preprocessing), the short-term hypertension score generator can perform short-term hypertension scoring on the input temporal tensor using at least the minimum number of pulse windows. In some examples, the minimum number of pulse windows can be one pulse window. In some examples, the minimum number of pulse windows can be greater than one pulse window (e.g., 2, 5, 10, etc.).

[0044] In some examples, the preprocessing circuit 455 can scale the pulse window. In some examples, the optical and / or motion data of each pulse window within the pulse window used for qualified pulses can be scaled by a channel-specific standard deviation (e.g., the first channel of optical data can be scaled by the standard deviation of the timing input tensor of the first channel). In some examples, the optical and / or motion data of each pulse window within the pulse window can be limited to a maximum value (e.g., a limiting value with an absolute value greater than 1).

[0045] The output of preprocessing circuit 455 (e.g., one or more filtered and scaled pulse windows) can be used as input to machine learning processing circuit 460. Machine learning processing circuit 460 can be a convolutional neural network (CNN). In some examples, the backbone of the CNN can consist of convolutional layers organized into residual blocks that form a “feature extractor” for a short-term hypertension score generator. The CNN can extract a set of features (short-term hypertension parameters) for each pulse window, which can be referred to herein as a feature vector. In some examples, the CNN can branch into a prediction head to generate feature representations of systolic and diastolic hypertension parameters, which can also be referred to as systolic feature vectors and diastolic feature vectors, respectively. In some examples, systolic and diastolic hypertension parameters / feature vectors can be combined into a single set of features (e.g., concatenating systolic and diastolic hypertension parameters into a single vector).

[0046] In some examples, the aggregation can be performed across pulse windows in the input (e.g., after generating the feature set for each pulse window in the input segment). In some examples, the aggregation can be the average of each feature in the feature set for each pulse window. In some examples, the aggregation can be the average of each feature in the systolic feature vector across the pulse windows in the input segment and the average of each feature in the diastolic feature vector across the pulse windows in the input segment.

[0047] In some examples, the feature set can branch into a prediction head, which performs computations on the feature set (e.g., aggregated systolic and diastolic feature vectors) to generate systolic and diastolic hypertension scores that can be correlated with the blood pressure of the input segment. The systolic and diastolic hypertension scores can then be used to generate a short-term hypertension score for the input segment. In some examples, one or more transformations can be used to implement the computation of generating the systolic and diastolic hypertension scores and the short-term hypertension score.

[0048] For example, such as Figure 4BAs shown, the feature representation output of CNN 460 can be transformed using transformation circuit 465 to apply one or more transformations to the feature representation to generate a short-term hypertension score 470. The short-term hypertension score 470 can be stored in memory. In some examples, transformation circuit 465 can apply one or more linear transformations to convert the high-dimensional feature vector output by CNN 460 into a scalar-valued short-term hypertension score. For example, a linear transformation can rotate the feature representation vector to a new basis. For example, a linear transformation can be used to transform the feature representation into mutually independent and orthogonal representations ordered by importance (e.g., the variance of features in the hypertension outcome). For example, principal component analysis (PCA) can be applied to the training data to learn the basis transformation matrix W that rotates the feature vectors into orthogonal representations. PCA One or more additional linear transformations can be applied to predict long-term systolic and diastolic blood pressure scores and / or long-term hypertension scores. For example, a first multi-output ridge regression (e.g., using L2 regularization) can be applied to the training data to predict long-term systolic and diastolic blood pressure scores on a new basis. A second ridge regression (e.g., using L2 regularization) can be applied to the training data to predict long-term hypertension status / scores from the predicted systolic and diastolic blood pressure scores. The regression weights obtained from the first and second ridge regressions can be used to learn a matrix W. BP (Systolic blood pressure weight / Diastolic blood pressure weight) and matrix W HT (Hypertension weighting)

[0049] In some examples, the linear transformation described above can be applied across multiple operations. For instance, matrix W can be used. PCA The first transformation is applied to the short-term eigenvectors to change the basis, and then W can be used. BP The second transformation is applied to the eigenvectors in the new basis to predict systolic and diastolic blood pressure scores, which can then be used with W. HT The third transformation is applied to the predicted systolic and diastolic blood pressure scores to predict short-term hypertension scores (e.g., scalar values). In some examples, some or all transformations can be combined and applied in fewer steps or a single step. For example, a single omnibus weight matrix W O It can be applied to the feature vector output by CNN 460 in a single transformation operation, where W O =W PCA *W BP *W HT A single composite weight can reduce storage requirements (one matrix instead of three) and processing time / operations (one transformation instead of three).

[0050] In some examples, the extracted feature data can be combined with systolic and diastolic hypertension parameters / feature vectors to form a combined feature set (e.g., concatenating the extracted feature data with systolic and diastolic hypertension parameters / vectors into a single vector). In some examples, the combination can occur before any transformation is performed. In some examples, the combination can occur after applying a first transformation (e.g., a first linear transformation to change the basis), and subsequent transformations (e.g., a second and third linear transformation) can be applied to the combined feature set. The extracted feature data can refine short-term hypertension scores and provide improved accuracy in hypertension estimation.

[0051] Although the summation of the feature set across the pulses described above applies to both the systolic and diastolic feature vectors, it should be understood that the summation can be applied at different processing stages. In some examples, the summation can be applied to the entire feature set earlier in the processing, before branching to individual systolic and diastolic heads. In some examples, the aforementioned systolic and diastolic feature vectors can be used to calculate the systolic and diastolic hypertension scores for each pulse's systolic and diastolic head, and then the systolic and diastolic hypertension scores can be summed across each pulse window in the input segment.

[0052] Figures 5A to 5C An exemplary block diagram of a long-term hypertension score generator according to an example of this disclosure is shown. Long-term hypertension score generator 500, long-term hypertension score generator 530, or long-term hypertension score generator 550 may correspond to... Figure 2 The long-term hypertension score generator 215 is described above. It should be understood that long-term hypertension score generators 500, 530, and 550 are exemplary implementations, and other implementations are also possible. More generally, the long-term hypertension score generator can receive short-term hypertension scores / parameters and output a long-term hypertension score (e.g., using a summation of short-term hypertension scores / parameters and / or using a machine learning model).

[0053] refer to Figure 5A The long-term hypertension score generator 500 may include a feature extraction block 505, a diastolic decision tree 510, and a systolic decision tree 515. The long-term hypertension score generator 500 can be used to generate a single long-term hypertension score using the aggregated short-term hypertension scores output by the short-term hypertension score generator 400 (e.g., a variable-length time series of short-term hypertension scores).

[0054] Feature extraction block 505 may receive short-term hypertension scores (e.g., stored in memory 210) as input. In some examples, feature extraction block 505 may extract statistical features from aggregated short-term hypertension scores. For example, the distribution of hypertension scores may be summarized by some or all of the mean, median, mode, variance, and / or percentiles, as well as other possible aggregate statistical measures.

[0055] The diastolic decision tree 510 and the systolic decision tree 515 can each be a gradient boosting decision tree machine learning model. The diastolic and systolic decision trees 510 and 515 can each receive the output of feature extraction block 505 as input and can output long-term systolic hypertension scores (which are correlated with total systolic blood pressure) and long-term diastolic hypertension scores (which are correlated with total diastolic blood pressure). The decision trees can be trained using short-term hypertension scores and long-term user-level blood pressure labels. Training can minimize the MAE between the long-term user-level blood pressure labels and the decision tree outputs. In some examples, to prevent overfitting, the number of trees that the gradient boosting decision tree can learn can be limited (e.g., based on error measurements on a validation dataset similar to but separate from the training data). Gradient boosting decision trees can learn different weighted parameters for the input features (each subsequent tree in the sequence corrects the errors of the preceding tree by applying different weights), allowing the ensemble of decision trees to provide a non-linear prediction function.

[0056] In some examples, such as Figure 5A As shown, the long-term hypertension score generator 500 may include separate gradient boosting decision trees to utilize feature combinations that may be unique for either systolic or diastolic blood pressure. The results (e.g., using a weighted average) can then be aggregated into a single long-term hypertension score 520, which can be used to estimate hypertension and / or report hypertension (e.g., as described with respect to process 300) when the long-term hypertension score exceeds a threshold. In some examples, instead of using separate systolic and diastolic gradient boosting decision trees 510 and 515, a set of gradient boosting decision trees can be used to generate the long-term hypertension score 520 (without interfering with the systolic and diastolic scores).

[0057] refer to Figure 5B The long-term hypertension score generator 530 may include a feature extraction block 535 and a machine learning model 540 (e.g., a regularized linear regression model). The long-term hypertension score generator 530 can be used to generate a single long-term hypertension score from short-term hypertension parameters aggregated by the short-term hypertension score generator 400. The feature extraction block 535 may receive short-term hypertension parameters (e.g., stored in memory 210) as input.

[0058] In some examples, the feature extraction block 535 may aggregate short-term hypertension parameters. For example, an aggregate statistical value of the short-term hypertension parameters may be calculated. In some examples, the aggregate statistical value may include a mean vector that is calculated to average a set of short-term hypertension parameters (one set per segment). For example, each estimate of the short-term hypertension parameters from the short-term hypertension score generator may include N parameters (features) that can be represented by a vector. The N parameters from each of the M short-term estimates (for M segments) may be averaged to produce a mean vector with N parameters (e.g., for N parameters, the average parameter 1 from each of the M short-term estimates, the average parameter 2 from each of the M short-term estimates, etc.). In some examples, the aggregate statistical value may also include the variance or standard deviation of each of the N parameters across the M short-term estimates. In some examples, other aggregate statistical values may be calculated.

[0059] In some examples, the feature extraction block 535 may also calculate the covariance (matrix) of the short-term hypertension parameters (e.g., the N parameters from each of the M short-term estimates). The covariance matrix may be represented by its eigenvectors and may be used as an additional parameter input to the machine learning model 540. In some examples, to reduce the number of input parameters, a smaller dimension may be used to estimate the covariance and / or the features may be represented by fewer eigenvectors of the covariance matrix. For example, in some examples, the short-term hypertension parameters may first be classified based on the variance of each of the N parameters (e.g., using principal component analysis), and the covariance may be calculated for a subset of the dimensions of the short-term hypertension parameters (for those parameters with the highest variance or within a range of variances < N dimensions). In some examples, an approximation of the covariance matrix may use a subset of the eigenvectors of the covariance matrix (e.g., one or more eigenvectors). This subset of eigenvectors may be used as an input to the machine learning model 540 together with the mean vector.

[0060] The machine learning model ********** may be a linear regression machine learning model (e.g., ridge regression). The machine learning models 540 may each receive the output of the feature extraction block 535 as input and may output a long-term hypertension score. The machine learning models may be trained using the short-term hypertension parameters (and associated extracted features) and the long-term user-level hypertension labels. The training may be used to minimize the MAE between the long-term user-level hypertension labels and the output of the machine learning model 540.

[0061] See Figure 5C It should be noted that there is an unclear expression "**********" in the content of ID=7, which may need to be further clarified according to the actual situation.The long-term hypertension score generator 550 can be used to generate a single long-term hypertension score using the aggregated short-term hypertension scores (e.g., a variable-length time series of short-term hypertension scores) output by the short-term hypertension score generator 450. The long-term hypertension score generator 550 may include an averaging block 555 to calculate the arithmetic mean of the short-term hypertension scores (e.g., output by the short-term hypertension score generator 450).

[0062] It should be understood that Figures 4A to 5C The components shown in the block diagram can be implemented in hardware or software, or a combination thereof. Furthermore, it should be understood that the block diagram is merely an example, and implementations may include fewer, more, or different blocks. For example, the filter of the short-term hypertension score generator 400 may be implemented in a separate part of the system (e.g., the filtered data stream may not be specifically used for hypertension monitoring). Additionally, it should be understood that components related to… Figures 5A to 5C Different aggregation techniques and / or feature extraction techniques and / or machine learning techniques are shown to generate long-term hypertension scores. For example, different aggregation techniques and / or feature extraction techniques and / or machine learning techniques can be used. Figure 5A Gradient boosting decision trees and Figure 5B Other machine learning models that use regularized linear regression models.

[0063] As described above, aspects of this technology include the collection and use of physiological information. This technology can be implemented in conjunction with technologies involving the collection of personal data related to a user's health and / or uniquely identifying or potentially used to contact or locate a specific person. Such personal data may include demographic data, date of birth, location-based data, telephone numbers, email addresses, home addresses, and data or records related to the user's health or health level (e.g., vital sign measurements, medication information, exercise information, etc.).

[0064] This disclosure recognizes that users' personal data (including physiological information, such as data generated and used by this technology) can be used to benefit users. For example, assessing a user's sleep patterns, heart rate, and / or blood pressure can allow the user to track or otherwise gain insights into their health.

[0065] This disclosure assumes that entities responsible for collecting, analyzing, disclosing, transmitting, storing, or otherwise using such personal data will comply with established privacy policies and / or privacy practices. Specifically, such entities should implement and adhere to privacy policies and practices that are recognized as meeting or exceeding industry or governmental requirements for maintaining the privacy and security of personal data. Such policies should be easily accessible to users and should be updated as data collection and / or use change. Personal information from users should be collected for the entity's lawful and reasonable purposes and not shared or sold outside of these lawful uses. Furthermore, such collection / sharing should require the user's informed consent. In addition, such entities should consider taking any necessary steps to protect and safeguard access to such personal data and ensure that others with access to personal data comply with their privacy policies and processes. Additionally, such entities may be subject to third-party evaluations to demonstrate their compliance with widely accepted privacy policies and practices. These policies and practices may be tailored to geographic regions and / or the specific type and nature of the personal data collected and used.

[0066] Regardless of the foregoing, this disclosure also envisions implementation schemes that allow users to selectively block the collection, use, or access to personal data, including physiological information. For example, users may be able to disable hardware and / or software components that collect physiological information. Additionally, this disclosure anticipates providing hardware and / or software components to prevent or block access to collected personal data. Specifically, users may choose to remove, disable, or restrict access to certain health-related applications that collect their personal health or fitness data.

[0067] Therefore, based on the foregoing, some examples of this disclosure relate to an electronic device. The electronic device may include: an optical sensor; a motion sensor; and processing circuitry coupled to the optical sensor and the motion sensor. The processing circuitry may be configured to: generate multiple estimates of a hypertension score or parameter, each corresponding estimate using a corresponding segment of data from the optical sensor and the motion sensor; and use the multiple estimates to generate a total hypertension score. As a supplement or alternative to one or more of the examples disclosed above, the processing circuitry may be further configured to: generate a notification about possible hypertension if the total hypertension score exceeds a threshold; and abandon the generation of the notification if the total hypertension score does not exceed the threshold. As a supplement or alternative to one or more of the examples disclosed above, in some examples, the corresponding segment may correspond to the duration of a first time period, and the total hypertension score may correspond to a second time period longer than the first time period. As a supplement or alternative to one or more of the examples disclosed above, in some examples, the processing circuitry may include a first machine learning model configured to generate multiple estimates of a hypertension score or parameter. As a supplement to or alternative to one or more of the examples disclosed above, in some examples, the first machine learning model may include a convolutional neural network. As a supplement to or alternative to one or more of the examples disclosed above, in some examples, the first machine learning model may include a first prediction head configured to generate a systolic hypertension score or parameter and a second prediction head configured to generate a diastolic hypertension score or parameter. As a supplement to or alternative to one or more of the examples disclosed above, in some examples, the processing circuitry may include a second machine learning model configured to generate a total hypertension score. As a supplement to or alternative to one or more of the examples disclosed above, in some examples, the second machine learning model may include one or more gradient boosting decision trees or regularized linear regression models. As a supplement to or alternative to one or more of the examples disclosed above, in some examples, generating a total hypertension score may include using multiple estimates to calculate statistical parameters and using the statistical parameters to generate the total hypertension score.

[0068] As a supplement or alternative to one or more of the examples disclosed above, in some examples, the processing circuitry may be further configured to divide corresponding segments of data from the optical sensor and the motion sensor into one or more pulse windows. As a supplement or alternative to one or more of the examples disclosed above, in some examples, the processing circuitry is further configured to scale one or more pulse windows. As a supplement or alternative to one or more of the examples disclosed above, in some examples, the processing circuitry may include a machine learning model configured to generate multiple estimates of a hypertension score or parameter. As a supplement or alternative to one or more of the examples disclosed above, in some examples, generating multiple estimates of a hypertension score or parameter using corresponding segments of data from the optical sensor and the motion sensor may include: inputting multiple pulse windows into the machine learning model to generate a feature vector of the hypertension parameter for each of the multiple pulse windows; and averaging the feature vectors of the multiple pulse windows to generate a total feature vector for the corresponding segment. As a supplement or alternative to one or more of the examples disclosed above, in some examples, generating multiple estimates of hypertension scores or parameters using corresponding segments of data from optical and motion sensors may include transforming the aggregate eigenvector of the corresponding segments to generate corresponding estimates with scalar values. As a supplement or alternative to one or more of the examples disclosed above, transforming the aggregate eigenvector may include applying one or more linear transformations. As a supplement or alternative to one or more of the examples disclosed above, in some examples, one or more linear transformations may include a transformation for changing the basis of the aggregate eigenvector of the corresponding segments to a new basis. As a supplement or alternative to one or more of the examples disclosed above, in some examples, one or more linear transformations may include a transformation for predicting systolic hypertension scores or parameters and diastolic hypertension scores or parameters based on the aggregate eigenvector of the corresponding segments in the new basis. As a supplement or alternative to one or more of the examples disclosed above, in some examples, one or more linear transformations may include a transformation for predicting corresponding estimates of hypertension scores based on systolic hypertension scores or parameters and diastolic hypertension scores or parameters. As a supplement to or alternative to one or more of the examples disclosed above, in some examples, generating the total hypertension score involves averaging multiple estimates to generate the total hypertension score. As a supplement to or alternative to one or more of the examples disclosed above, in some examples, features extracted from the optical data may be added to the total feature vector before or during one or more linear transformations.

[0069] Some examples of this disclosure relate to a method. The method may include: generating multiple estimates of a hypertension score or parameter, each of the multiple estimates using a corresponding segment of data from an optical sensor and a motion sensor; and using the multiple estimates to generate a total hypertension score. As a supplement or alternative to one or more of the examples disclosed above, in some examples, the method may further include: generating a notification about possible hypertension based on the total hypertension score exceeding a threshold; and abandoning the generation of the notification based on the total hypertension score not exceeding the threshold. As a supplement or alternative to one or more of the examples disclosed above, in some examples, the corresponding segment may correspond to the duration of a first time period, and the total hypertension score may correspond to a second time period longer than the first time period. As a supplement or alternative to one or more of the examples disclosed above, in some examples, generating multiple estimates of a hypertension score or parameter may include applying a first machine learning model to multiple segments of data from the optical sensor and the motion sensor. As a supplement or alternative to one or more of the examples disclosed above, in some examples, the first machine learning model may include a convolutional neural network. As a supplement to or alternative to one or more of the examples disclosed above, in some examples, the first machine learning model may include a first predictive head configured to generate a systolic hypertension score or parameter and a second predictive head configured to generate a diastolic hypertension score or parameter. As a supplement to or alternative to one or more of the examples disclosed above, in some examples, generating a total hypertension score may include applying a second machine learning model to multiple estimates. As a supplement to or alternative to one or more of the examples disclosed above, in some examples, the second machine learning model may include one or more gradient boosting decision trees or regularized linear regression models. As a supplement to or alternative to one or more of the examples disclosed above, in some examples, generating a total hypertension score may include using multiple estimates to compute statistical parameters and using the statistical parameters to generate a total hypertension score.

[0070] As a supplement or alternative to one or more of the examples above, in some examples, the method may further include dividing corresponding segments of data from the optical sensor and motion sensor into one or more pulse windows. As a supplement or alternative to one or more of the examples disclosed above, in some examples, the method may further include scaling one or more pulse windows. As a supplement or alternative to one or more of the examples above, in some examples, generating multiple estimates of a hypertension score or parameter may include applying a machine learning model configured to generate multiple estimates of a hypertension score or parameter. As a supplement or alternative to one or more of the examples disclosed above, in some examples, using corresponding segments of data from the optical sensor and motion sensor to generate corresponding estimates of a hypertension score or parameter may include: inputting multiple pulse windows into a machine learning model to generate a feature vector of a hypertension parameter for each of the multiple pulse windows; and averaging the feature vectors of the multiple pulse windows to generate a total feature vector for the corresponding segment. As a supplement or alternative to one or more of the examples disclosed above, in some examples, generating multiple estimates of hypertension scores or parameters using corresponding segments of data from optical and motion sensors may include transforming the aggregate eigenvector of the corresponding segments to generate corresponding estimates with scalar values. As a supplement or alternative to one or more of the examples disclosed above, transforming the aggregate eigenvector may include applying one or more linear transformations. As a supplement or alternative to one or more of the examples disclosed above, in some examples, one or more linear transformations may include a transformation for changing the basis of the aggregate eigenvector of the corresponding segments to a new basis. As a supplement or alternative to one or more of the examples disclosed above, in some examples, one or more linear transformations may include a transformation for predicting systolic hypertension scores or parameters and diastolic hypertension scores or parameters based on the aggregate eigenvector of the corresponding segments in the new basis. As a supplement or alternative to one or more of the examples disclosed above, in some examples, one or more linear transformations may include a transformation for predicting corresponding estimates of hypertension scores based on systolic hypertension scores or parameters and diastolic hypertension scores or parameters. As a supplement to or alternative to one or more of the examples disclosed above, in some examples, generating the total hypertension score involves averaging multiple estimates to generate the total hypertension score. As a supplement to or alternative to one or more of the examples disclosed above, in some examples, features extracted from the optical data may be added to the total feature vector before or during one or more linear transformations.

[0071] Some examples of this disclosure relate to non-transitory computer-readable storage media. Non-transitory computer-readable storage media may store instructions that, when executed by a device including processing circuitry, may cause the processing circuitry to perform any of the methods described above.

[0072] Although examples of this disclosure have been fully described with reference to the accompanying drawings, it should be noted that various changes and modifications will become apparent to those skilled in the art. It should be understood that such changes and modifications are considered to be included within the scope of the examples of this disclosure as defined by the appended claims.

Claims

1. An electronic device, comprising: Optical sensors; Motion sensor; as well as A processing circuit, coupled to the optical sensor and the motion sensor, is configured to: Generate multiple estimates of a hypertension score or a parameter indicating a hypertension score, each of the multiple estimates of the hypertension score or the parameter indicating the hypertension score being generated using a corresponding segment of data from the optical sensor and the motion sensor within a first duration of a first time period, wherein the first duration of the first time period is a part of a day; The multiple estimates are used to generate a total hypertension score, wherein the total hypertension score indicates the hypertension score within a second duration of a second time period, wherein the second duration of the second time period is two days or more. If the total hypertension score exceeds a threshold, a notification about possible hypertension is generated and the notification is output to a health application or one or more output devices. as well as If the total hypertension score does not exceed the threshold, the generation of the notification is abandoned and the output of the notification to the health application or the one or more output devices is abandoned.

2. The electronic device of claim 1, wherein the processing circuitry includes a first machine learning model configured to generate a hypertension score or estimates of parameters indicating the hypertension score.

3. The electronic device of claim 2, wherein the first machine learning model comprises a convolutional neural network.

4. The electronic device of claim 3, wherein the first machine learning model includes a first prediction head configured to generate a systolic hypertension score or parameters indicating the systolic hypertension score and a second prediction head configured to generate a diastolic hypertension score or parameters indicating the diastolic hypertension score.

5. The electronic device of claim 2, wherein the processing circuitry includes a second machine learning model configured to generate the total hypertension score.

6. The electronic device of claim 5, wherein the second machine learning model comprises one or more gradient boosting decision trees or regularized linear regression models.

7. The electronic device of claim 1, wherein generating the total hypertension score comprises: The multiple estimates are used to calculate statistical parameters, and the statistical parameters are used to generate the total hypertension score.

8. The electronic device according to claim 1, wherein the processing circuit is further configured to: The corresponding segments of data from the optical sensor and the motion sensor are divided into one or more pulse windows.

9. The electronic device according to claim 8, wherein the processing circuit is further configured to: Scaling the one or more pulse windows.

10. The electronic device of claim 8, wherein the processing circuitry includes a machine learning model configured to generate a hypertension score or estimates of parameters indicative of the hypertension score.

11. The electronic device of claim 10, wherein using the corresponding segments of data from the optical sensor and the motion sensor to generate a hypertension score or a parameter indicating the hypertension score comprises: Multiple pulse windows are input into the machine learning model to generate a feature vector of hypertension parameters for each of the multiple pulse windows; as well as The feature vectors of the multiple pulse windows are averaged to generate the sum of the feature vectors of the corresponding segments.

12. The electronic device of claim 11, wherein using the corresponding segments of data from the optical sensor and the motion sensor to generate a hypertension score or a parameter indicating the hypertension score comprises: Transform the aggregate eigenvector of the corresponding fragments to generate the corresponding estimate with a scalar value.

13. The electronic device of claim 12, wherein transforming the aggregate feature vector comprises applying one or more linear transformations.

14. The electronic device of claim 13, wherein the one or more linear transformations include a transformation for changing the basis of the aggregate eigenvectors of the respective segments to a new basis.

15. The electronic device of claim 14, wherein the one or more linear transformations include transformations for predicting systolic hypertension scores or parameters indicative of systolic hypertension scores and diastolic hypertension scores or parameters indicative of diastolic hypertension scores from the aggregate eigenvectors of the respective segments in the new basis.

16. The electronic device of claim 15, wherein the one or more linear transformations include a transformation for predicting the corresponding estimates of the hypertension score from the systolic hypertension score or a parameter indicating the systolic hypertension score and the diastolic hypertension score or a parameter indicating the diastolic hypertension score.

17. The electronic device of claim 1, wherein generating the total hypertension score comprises averaging the plurality of estimates to generate the total hypertension score.

18. A method comprising: Multiple estimates of a hypertension score or a parameter indicating a hypertension score are generated, each of the multiple estimates of the hypertension score or the parameter indicating the hypertension score being generated using a corresponding segment of data from an optical sensor and a motion sensor within a first duration of a first time period, wherein the first duration of the first time period is a part of a day; The multiple estimates are used to generate a total hypertension score, wherein the total hypertension score indicates the hypertension score within a second duration of a second time period, wherein the second duration of the second time period is two days or more. If the total hypertension score exceeds a threshold, a notification about possible hypertension is generated and the notification is output to a health application or one or more output devices. as well as If the total hypertension score does not exceed the threshold, the generation of the notification is abandoned and the output of the notification to the health application or the one or more output devices is abandoned.

19. The method of claim 18, wherein generating the plurality of estimates of a hypertension score or parameters indicating a hypertension score comprises applying a first machine learning model to a plurality of segments of data from the optical sensor and the motion sensor.

20. The method of claim 19, wherein the first machine learning model comprises a convolutional neural network.

21. The method of claim 20, wherein the first machine learning model includes a first prediction head configured to generate a systolic hypertension score or parameters indicating the systolic hypertension score and a second prediction head configured to generate a diastolic hypertension score or parameters indicating the diastolic hypertension score.

22. The method of claim 19, wherein generating the aggregate hypertension score comprises applying a second machine learning model to the plurality of estimates.

23. The method of claim 22, wherein the second machine learning model comprises one or more gradient boosting decision trees or regularized linear regression models.

24. The method of claim 18, wherein generating the total hypertension score comprises: The multiple estimates are used to calculate statistical parameters, and the statistical parameters are used to generate the total hypertension score.

25. The method of claim 18, further comprising: The corresponding segments of data from the optical sensor and the motion sensor are divided into one or more pulse windows.

26. The method of claim 25, further comprising: Scaling the one or more pulse windows.

27. The method of claim 25, wherein generating the plurality of estimates of a hypertension score or parameter comprises applying a machine learning model to generate the plurality of estimates of a hypertension score or a parameter indicating the hypertension score.

28. The method of claim 27, wherein using the corresponding segments of data from the optical sensor and the motion sensor to generate a hypertension score or a corresponding estimate of a plurality of estimates of a parameter indicating the hypertension score comprises: Multiple pulse windows are input into the machine learning model to generate a feature vector of hypertension parameters for each of the multiple pulse windows; as well as The feature vectors of the multiple pulse windows are averaged to generate the sum of the feature vectors of the corresponding segments.

29. The method of claim 28, wherein using the corresponding segments of data from the optical sensor and the motion sensor to generate a hypertension score or a corresponding estimate of a plurality of estimates of a parameter indicating the hypertension score comprises: Transform the aggregate eigenvector of the corresponding fragments to generate the corresponding estimate with a scalar value.

30. The method of claim 29, wherein transforming the aggregate feature vector comprises applying one or more linear transformations.

31. The method of claim 30, wherein the one or more linear transformations include a transformation for changing the basis of the aggregate eigenvectors of the respective segments to a new basis.

32. The method of claim 31, wherein the one or more linear transformations include transformations of parameters for predicting systolic hypertension scores or indicative of systolic hypertension scores and diastolic hypertension scores or indicative of diastolic hypertension scores from the aggregate eigenvectors of the corresponding segments in the new basis.

33. The method of claim 32, wherein the one or more linear transformations include a transformation for predicting the corresponding estimates of the hypertension score from the systolic hypertension score or a parameter indicating the systolic hypertension score and the diastolic hypertension score or a parameter indicating the diastolic hypertension score.

34. The method of claim 18, wherein generating the total hypertension score comprises averaging the plurality of estimates to generate the total hypertension score.

Images